Last summer, I wrote a post setting out how ineffective antidepressant treatment for my depression had been. In the second half of the post, I went on to explain why I thought antidepressants were likely to always be ineffective for me. That part of the post was necessarily more speculative, so I was intrigued to discover some evidence which suggests that, against all expectations, I may have actually been saying something relevant, as opposed to just chuntering on in my usual ignorant fashion.
Let me start by reproducing (in a not at all narcissistic way) the relevant part of my original post.
Someone like me who, thanks to a combination of first-hand experience and reading around the subject, knows that any new antidepressant he tries will have, at the very best, only a minimal effect on his depression will inevitably start taking the drug with a pessimistic mindset, an expectation that this drug will fail to lift his mood, just like all the others did. […] having so negative a mindset means that the very fragile part of the response to the drug that is not a result of the placebo effect – that very slight impact on mood – is at great risk of being overwhelmed and rendered invisible by the weight of negative expectation that surrounds it. It’s not just a question of the placebo effect failing to function, but of a kind of negative-placebo-effect taking its place.
I find this theory of mine – that my lack of belief in the efficacy of antidepressants is what makes them so particularly ineffective for me – very interesting in the light of some recently published research which was reported on the BBC News website.
So far as I can establish from the BBC report and the abstract (the paper itself is behind a paywall), volunteers were put in an MRI scanner, and had ‘constant heat pain’ applied to their leg. (It will always be a mystery to me why people volunteer for these experiments; there isn’t enough money in the world, surely?) They were also hooked up to an IV drip, which meant they could be dosed with pain medication without realising it. The test subjects were asked to rate the level of pain they were experiencing on a scale of one to a hundred, and the activation of the areas of their brains associated with the perception of pain was also assessed.
In the first part of the test, subjects were given no pain medication; in the second part, pain medication was introduced without their knowledge; in the third, medication was maintained at the same dose, but subjects were told it was being introduced for the first time; and in the fourth stage, the dose was again maintained, but the subjects were told medication was being withdrawn, and that they should expect to feel pain. These were the results of the self-assessment of pain (manky little table drawn up by me, based on the results as reported in the BBC article):
However, it also shows that the introduction of pain medication had a modest impact on the perception of pain (stage 2), but that this effect was greatly enhanced when the test subjects were told they were receiving pain medication (stage 3). So far this is a nice demonstration of the powerful impact of the placebo effect; the variation in pain between the two stages is entirely the result of the placebo effect, since medication was maintained at the same dose throughout. Where things become interesting is at stage 4, which demonstrates that (as near as damn it) all the benefit of pain medication was counteracted by the subjects’ belief that it had been withdrawn; it’s this that suggests the operation of an ‘anti-placebo’ effect even more powerful than the placebo effect. These self-assessment scores were also borne out by the MRI results, which showed significant changes in neural activity associated with pain between stages 2, 3 and 4 of the experiment, even though the dose of pain medication remained constant throughout.
On the basis of all this, the researchers say the following:
we contend that an individual’s expectation of a drug’s effect critically influences its therapeutic efficacy
This is one of those times when I’d dearly like access to the paper as a whole, because it strikes me that the experiment as reported in the abstract doesn’t entirely support that assertion. The experiment seems to me to have demonstrated that the therapeutic efficacy of a drug was greatly enhanced by subjects being told they were receiving it, and that its efficacy was almost totally destroyed by subjects being told it had been withdrawn. It strikes me that it was not the subjects’ belief in the efficacy of the drug that varied, but rather their understanding of whether or not they were receiving a drug at all. So far as it’s possible to tell from the data I have access to, it seems the participants must have believed that the drug was efficacious in order to have experienced such significant variations in pain when they were told the drug was being given and withdrawn.
It’s possible, of course, that this apparent confusion is cleared up in the full text of the paper (and just as possible that I’ve missed something of significance in the abstract, and that the confusion is all mine). Even if not, it’s still the case that the experiment does seem to demonstrate the presence of a powerful ‘anti-placebo’ effect. I find this interesting, because it seems to intersect so neatly with my own anecdotal experience of antidepressants.
As I set out last June, my extensive experience with antidepressants has led to me being pretty much convinced that they’re ineffective – or, as I should say, ineffective for me – and this means that, if I were to start taking them again, my chances of experiencing a benefit would be largely wiped out by the weight of my negative expectations. This study seems to bear that out, and to be evidence in support of what I suggested then – that antidepressants can only help people who believe in them, and that, since I no longer believe, I can no longer be helped.
I guess I should be careful about over-interpretation. This study was only carried out on 22 people, and it focussed on a completely different class of medication. Pain is also famously variable, and the effectiveness of pain medication was already known to be materially affected by all kinds of, for want of a better word, subjective factors. Then again, the experience of depression is, like pain, wholly generated by the mind, and is also known to be affected by all kinds of ‘subjective’ factors. In fact, if you asked for my inexpert opinion on which other class of medication these findings were most likely to apply to, antidepressants would probably be my answer, on the basis that, as with pain medication, there’s no truly objective way to measure their effectiveness. With antidepressants as with pain meds, the only measure of their success is how they make the person taking them feel, and there’s nothing more subjective than that.
Standard disclaimer: I am so not an expert on any of this. I don’t have any medical training whatsoever, and my highest scientific qualification is a grade D in A level biology. Please make sure you treat everything I say with the appropriate level of caution – which is to say lots and lots of caution. In fact, you’re probably best off pretending you’ve never read this…